Bacteriophages are the most numerous organisms in the biosphere and play major roles in the ecology, evolution, physiology and pathogenesis of their bacterial hosts. Their relative simplicity makes them terrific model systems and they have been instrumental in the development of molecular biology and the biotechnology industry. Nevertheless, notwithstanding the in-depth studies on a small number of bacteriophages, our understanding of phage diversity, evolution and genomic constitution is rudimentary at best. Elucidation of bacteriophage diversity and the evolutionary mechanisms that generate new viruses can be investigated by comparative sequence analysis of phage genomes. Analysis of the currently available sequences of approximately 85 genomes of dsDNA tailed phages reveals that they are extremely diverse at the nucleotide sequence level and that pairs of phages competent to infect a common host do not necessarily share any significant DNA sequence homology. When analyzed at the amino acid sequence level it is apparent that while there are many genes shared among phages, there are also a high proportion of previously unidentified sequences. Bacteriophages therefore constitute a huge reservoir of unexplored sequence information. Bacteriophage genomes are characteristically mosaic, composed of modules that appear to migrate throughout the phage population. This mosaicism is generated by a combination of illegitimate and homologous recombination, with novel mosaic boundaries generated by illegitimate events, which are then reassorted by homologous recombination. Comparative sequence analysis reveals insights into how these events occur at the genetic and molecular levels. The total number of sequenced phage genomes is still sufficiently small that it is not possible to address population level questions, such as which genomes are participating in recombination and what is the direction and rate of gene flow between phages and between phages and their hosts. A larger set of phage genomic information of viruses that infect more closely-related bacterial hosts and of viruses with extremely large genomes (approximately 500kbp) will reveal how phages evolve, how they influence the evolution of their hosts, and what - at the most fundamental level - distinguishes viral and host genomes.